Marginal Emission Factors Considering Renewables: A Case Study of

Humphrey School of Public Affairs, University of Minnesota, 301 19th Avenue South,. 10 ... technology mix, fuel type, operation cost, dispatchability,...
1 downloads 0 Views 861KB Size
Subscriber access provided by GRIFFITH UNIVERSITY

Article

Marginal Emission Factors Considering Renewables: A Case Study of the U.S. Midcontinent Independent System Operator (MISO) System Mo Li, Timothy Michael Smith, Yi Yang, and Elizabeth J Wilson Environ. Sci. Technol., Just Accepted Manuscript • DOI: 10.1021/acs.est.7b00034 • Publication Date (Web): 01 Sep 2017 Downloaded from http://pubs.acs.org on September 4, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Environmental Science & Technology is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29

30 31 32 33 34 35 36 37 38

Marginal Emission Factors Considering Renewables: A Case Study of the U.S. Midcontinent Independent System Operator (MISO) System Mo Li1, Timothy M. Smith1*, Yi Yang2, Elizabeth J. Wilson3 1

NorthStar Initiative for Sustainable Enterprise, Institute on the Environment, University of Minnesota, 325 Learning and Environmental Sciences, 1954 Buford Ave, Saint Paul, Minnesota 55108, United States 2 CSRA Inc., 3170 Fairview Park Drive, Falls Church, VA 22042, United States 3 Humphrey School of Public Affairs, University of Minnesota, 301 19th Avenue South, Minneapolis, MN 55455, United States Abstract Previous marginal emission factors (MEFs) estimate for the electricity sector have focused on emitting sources only, assuming that non-emitting renewables rarely contribute to marginal generation. However, with increased penetration and improved dispatch of renewables, this assumption may be outdated. Here, we improve the methodology to incorporate renewables in MEF estimates and demonstrate a case study for the Midcontinent Independent System Operator (MISO) system where wind has been commonly dispatched on the margin. We also illustrate the spatiotemporal variations of MEFs and explore implications for energy storage technologies. The results show that because the share of renewables in MISO is still fairly low (6.34%), conventional MEFs focused on emitting sources provide a good estimate in MISO overall, as well as in the Central and South sub-regions. However, in the MISO North sub-region where wind provides 22.5% of grid generation, neglecting non-emitting sources can overestimate the avoided CO2, SO2, and NOX emissions by about 30%. Our study highlights the importance of considering renewables in MEFs in regions with high renewables penetration, and suggests energy policies to account for their impacts in these regions.

Keywords: Expanded Marginal Emission Factors (MEFs), Dispatchable Intermittent Resources (DIRs), Electricity Generation, Energy Policy 1. Introduction Electricity generation is one of the most emission-intensive economic sectors in the United States (U.S.) and across the globe.1 Regional air pollution and its consequential damage to human health are largely attributable to air pollutant emissions from fossil fuel combustion in 1 ACS Paragon Plus Environment

Environmental Science & Technology

Page 2 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60 61 62 63 64 65 66 67 68 69 70 71 72 73 74 75 76 77 78 79 80 81 82 83 84

electricity generation.2,3 Climate, air pollution, and energy policies aim to create more sustainable energy systems and reduce emissions from electricity generation. Commonly suggested policy interventions include increasing electricity production from renewable energy sources like wind and solar, decreasing consumption by improving energy efficiency, and shifting demand by using bulk energy storage and demand response.4,5 Effective implementation of these policies depends on understanding how they change and affect the electricity generation system and contribute to reducing emissions. The potential of an electricity generation system to reduce emissions has been measured by both marginal emission factors (MEFs) and average emission factors (AEFs).6–8 However, the use of AEFs, which reflect grid-average situations, to estimate the effect of an invention may be problematic, because not all generating technologies would respond to changes in demand proportionally.6 In studying the electricity generation in the United States, for example, SilerEvans et al. found that AEFs could significantly misestimate the amount of emissions avoided by an intervention.7 By contrast, MEFs estimate the emission intensity of marginal power generation that responds to a change in demand, and are a more appropriate metric to assess emission implications of policy and technology interventions, such as electric cars tax credit and energy storage.6–8 Short-term MEFs reflect the dynamics of electric generation and consumption given relatively fixed and long-lived system capacity; these are affected by factors including the legacy technology mix, fuel type, operation cost, dispatchability, and timing.9 Long-term—which in the electric power system can be decades—MEFs reflect capacity addition or reduction and structural changes in the electricity system, and are affected by factors such as resource constraints, capital cost and return rate, and policy incentives.9 Short-term MEFs have been intensively studied, mainly via two approaches: 1) statistical models based on empirical data and 2) economic dispatch models.10 Both methods have strengths and limitations.10 The statistical approach reduces model complexity and calculation time by avoiding operational constraints and dispatch orders, but it relies heavily on empirical data.6,7,10 When important data are missing, the accuracy of this approach is compromised.7 The economic dispatch approach estimates MEFs based on numerous data and sophisticated models.10–12 However, its increased complexity and strict assumptions confine broader use.13–15 Regressions on historical data have been used as a simple and effective statistical method of estimating MEFs. Siler-Evans et al. (2012) developed a linear regression model to calculate MEFs for eight North American Electric Reliability Corporation (NERC) regions using hourly, generator-level emissions and generation data from the U.S. Environmental Protection Agency (EPA) Continuous Emissions Monitoring System (CEMS).7 Holland and Mansur (2008) determined real-time pricing’s effect on emissions in the NERC regions.16 In Europe, Hawkes (2010) regressed and estimated marginal CO2 rates for the United Kingdom (UK) national grid from 2002 through 2009.6 The MEFs mentioned above have been further refined considering spatial and temporal variations, and have been broadly used to evaluate emissions associated with plug-in hybrid electric vehicles (PHEVs).17–24 The U.S. EPA also provides non-baseload MEFs in the Emissions & Generation Resource Integrated Database (eGRID);25 the method and results have been widely used in governmental policies and tools such as EPA’s Clean Power Plan (CPP) and Power Profiler application.26 Economic dispatch models have been used to estimate MEFs in different regional grids in the US: California Independent System Operator (CAISO),27 Electric Reliability Council of Texas (ERCOT),28,29 Midcontinent Independent System Operator (MISO),30 New York Independent 2 ACS Paragon Plus Environment

Page 3 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104 105 106 107 108 109 110 111 112 113 114 115 116 117

System Operator (NYISO),28 and Pennsylvania-New Jersey-Maryland Interconnection (PJM);29 and Europe: the power system of England and Wales30 and the Belgian electricity generation system.31 MEFs implications have been intensively examined in various generation- and demand-side applications, including PHEVs,11–15,32–36 integration of renewable energy,37,38 distributed generation,39 energy efficiency,40,41 energy storage,5,42–45 and a ban on new coal-fired power plants.46 One critical commonality of the existing studies is the exclusion of non-emitting generation from the calculation of MEFs, specifically renewable and nuclear sources. This is partly because of the lack of data7, but largely because of the assumption that non-emitting sources, particularly renewable resources, do not serve as marginal generation given their near-zero operational cost and non-dispatchability.6,7,18 However, in the U.S. significant deployment of renewables (over 75,000 MW in 2016) and parallel developments, including transmission system expansion, dispatching mechanism improvement, and electricity market development, in regional electricity systems, such as MISO and PJM, have shown that renewable resources like wind can be and have been deployed frequently on the margin (Figure 1).47,48 In 2014, for example, MISO’s monthly average percentage of time that wind was on the margin exceeded 60% in seven out of the twelve months.47 The remarkable presence of wind generation on the margin is due to technological and policy innovations affecting dispatch of MISO’s 16,000 MW of wind, in particular, the Dispatchable Intermittent Resources (DIRs) program launched in 2011.49,50 The program relies on advanced wind forecasting accuracy in MISO and enables automatic wind dispatchment and curtailment in real-time, with the aim of improving market efficiency and system reliability and reducing curtailments.49 In 2014, MISO’s 5-min interval forecast was developed and improved; MISO’s Day-Ahead and 4-Hour-Ahead hourly wind forecasting accuracy exceeded 95%;51 and monthly DIR participation in MISO accounted for 79%-83% of the total wind generation, a significant increase from 9.8% in June, 2011.49 Thanks to the DIRs program, wind participated in MISO’s real-time market like fossil fuels with a fairly low monthly curtailment rate ranging from 4.2% to 8.3% in 201452. This significantly improves wind’s dispatchability that indicates wind generation’s responsiveness to the request of power system operator (i.e. MISO) or power plant owner. Also, to better capture power system dynamics, the Independent System Operator New England (ISO-NE) has revised their estimates of MEFs.53 Their early MEFs covered only natural gas- and oil-fired generators, but from 2011 they have used a new method to estimate MEFs based on the locational marginal units, including non-emitting generators, identified by locational marginal prices.

3 ACS Paragon Plus Environment

Environmental Science & Technology

Page 4 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS

118 119 120 121 122 123 124 125 126 127 128 129 130 131 132 133 134 135 136 137 138 139 140 141 142 143 144 145 146 147 148

Figure 1. Monthly percentage of time that wind is on the margin in 2014, 2015, and 2016 in MISO. Data are from MISO’s 5-min real-time fuel on the margin data.54–56 Note that more than one fuel type may be on the margin, hence the percentage of each fuel type may sum to more than 100%.

Recognizing the recent developments in renewable penetration and policy innovations, here we apply the broadly-used linear regression approach6,7,18 to estimate MEFs for MISO, taking into account both emitting and non-emitting sources. We term our estimates “expanded MEFs” to differentiate from previous MEFs, which we term “conventional MEFs”. Because of MISO’s largest geographic footprint in the U.S.57 (900,000 square miles or 2.3 million square kilometers) and the substantial heterogeneity in fuel mix across its regions,58 we estimate MEFs for its North, Central, and South sub-regions, as well as MISO as a whole. We compare the differences between the expanded MEFs and the conventional MEFs to assess the impact of non-emitting sources such as renewables on MEFs. We also demonstrate spatiotemporal variations of the expanded MEFs and explore the implications for bulk energy storage technologies and demand side management (DSM) programs that are based on load shifting techniques such as demand response (DR), smart metering, and other emerging technologies. In addition, we inspect the fuel mix in marginal generation with respect to system load. We conclude with discussion of the broad policy and technology implications of MEFs considering non-emitting sources. 2. Methods We estimate MEFs using the linear regression approach6,7 with 2014 hourly emissions and generation data. Based on this approach, changes in emissions (∆ ) are a function of changes in generation (∆ ) within an hour (Equation 1): ∆  ∆  

(1)

where ∆   

(2)

and ∆   

(3)

4 ACS Paragon Plus Environment

Page 5 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194

The slope of the linear regression () gives us an estimate of MEFs. In the previous MEFs studies, ∆ covers only emitting sources (Equation 4): ∆   ,  ,

(4)

When the penetration of renewables is low or they are undispatchable and thus accounting for emitting sources only may be adequate for the estimation of MEFs. However, when renewables penetration becomes high with substantially improved dispatchability, accounting for emitting sources only may significantly overestimate the MEFs. In our expanded MEFs (Equation 5), we include both emitting and non-emitting sources, ∆  ( ,   , ) ( ,   , )

(5)

We use the approach to separately calculate MEFs of electricity generation in MISO’s North, Central, and South sub-regions. With regard to geographical coverage, North includes Iowa, Minnesota, Montana, North Dakota, South Dakota, and Manitoba, Canada; Central includes Indiana, Illinois, Kentucky, Michigan, Missouri, and Wisconsin; and South includes Arkansas, Louisiana, Mississippi, and Texas.57 Details of the MISO sub-regions are shown in the supporting information (section 1). The approach can also be applied to subsets of the sample to derive MEFs with more granular temporal resolutions. For example, month-hour MEFs can be calculated by employing the regression on all observations within the same hourly interval in twelve months for 288 separate times. And daily MEFs can be calculated by employing the regression on all observations within the same hourly interval for 24 separate times. 2.1.Data Sources The primary data sources for our expanded MEFs estimates are the EPA’s Air Market Program Data (AMPD), which provides emissions and generation data for emitting sources, and hourly generation by fuel type data from MISO, which provides generation data for non-emitting sources. The AMPD is a web-based application that contains data collected as part of EPA’s emissions trading programs.59 The AMPD data provide hourly generator-level CO2, SO2 and NOX emissions as well as gross power production. We sort the data into the MISO sub-regions by cross-indexing power plant identification numbers with survey Form EIA-860, an independent database maintained by U.S. Energy Information Administration, and the MISO sub-region map.57,60 We then aggregate the hourly generator-level emissions and generation data to create the AMPD hourly sub-regional emissions and generation data. The hourly generation data provided by MISO covers all fuel types, but we only use the information on non-emitting sources, i.e. hydro, nuclear and wind. To accurately estimate the contribution of non-emitting sources in MEFs, we use an identification index to help determine what fuel types are on the margin in a specific hourly interval for each MISO sub-region. We collect the 5-minute real-time fuel on the margin data from MISO.54 In the data, MISO defines the marginal fuel(s) as the type(s) of generating units that are dispatched to serve the next 1 MW of energy61 for each fuel type, not the actual generation volume. We aggregate the data from 5minute to hourly level and obtain the total count of non-emitting generating units that operate on the margin within each hourly interval. The total count is then transformed into the identification index that indicates the presence or absence of particular non-emitting fuel types in marginal 5 ACS Paragon Plus Environment

Environmental Science & Technology

Page 6 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240

generation. Eventually, we use the index to filter the hourly non-emitting generation data, keeping the observations that are identified “on the margin”, and saving them as the MISO hourly non-emitting generation data for each MISO sub-region. Note that there can be one or more on-the-margin fuel types within each hourly interval. 2.2.Marginal Fuel Type We rely on the linear regression model and the MISO hourly generation by fuel type data to estimate each fuel type’s share in marginal generation across various temporal scales and grid production levels. To constrain our analysis more precisely to the identified on-the-margin generation, we use the identification index created from MISO’s 5-minute real-time fuel on the margin data to filter the hourly generation by fuel type data, then keep the observations that are identified “on the margin” as a subset. In this subset, we calculate change in generation for all fuel types within each hourly interval, then apply the similar linear regression model used for estimating MEFs to calculate the share of all possible fuel types in marginal generation. We also explore the share of fuel type in marginal generation with regard to system generation increase. We bin the hourly system generation by every fifth percentile and apply separate regressions on data within each bin. This allows us to intuitively observe the probability of each fuel type being on the margin at different levels of system generation, which is a proxy of system demand. 3. Results 3.1.Expanded and Conventional MEFs Overall results of the expanded MEFs and the conventional MEFs for MISO and its subregions, as well as marginal fuel type under the two methods are presented Table 1. MEFs estimates for CO2, SO2, and NOX are reported with ± two standard deviations of the estimates and R2 values (details of the linear regression are shown in SI section 2 and Figure S1). The difference between the two MEFs reflects the impact of non-emitting sources, primarily wind in this case. It also reflects the error of overestimation by using conventional MEFs when nonemitting sources contribute considerably to marginal electricity generation. Tables 1 reveals that the overall error is small for MISO, but it can be as high as around 30% for the North due to the high wind penetration (22.5% of 2014 grid generation58). The more wind is on the generation margin (North > Central > South), the greater the error of overestimation by using the conventional MEFs (e.g., 28.13% > 3.7% > 0.0% for CO2). Because the share of renewables like wind and hydro in MISO is still fairly low (less than 10%58, also shown in Table S1) and nuclear reactors rarely operate on the margin, estimates of the expanded and conventional MEFs do not show a difference in MISO overall, as well as in the Central and South sub-regions. The expanded MEFs for CO2 are found to be 0.76 tons/MWh for MISO in 2014, with a variation between 0.62 and 0.82 tons/MWh in different sub-regions. CO2 emissions are lower in the North (0.67 tons/MWh) and South (0.62 tons/MWh) than in the Central region (0.82 tons/MWh). This is because wind contributes 30% to marginal electricity generation in North and natural gas contributes nearly 80% in South. Our model’s R2 value is large in MISO (0.95), Central (0.96) and South (0.93), indicating that CO2 emissions respond strongly to changes in system generation. It also echoes the fact that coal and gas constitute the majority of marginal generation within these regions and sub-regions. The R2 value in the North regression is lower in North (0.64), indicating that higher wind penetration in marginal generation (30.4%) weakens the corrolation between changes in system generation and changes in CO2 emissions, because wind has zero emissions in operation. 6 ACS Paragon Plus Environment

Page 7 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262

SO2 is a major emission from coal combustion, therefore its expanded MEFs are higher in MISO (2.59 lbs/MWh), North (1.99 lbs/MWh) and Central (3.34 lbs/MWh) where marginal generation is more coal-dependent, but lower in South (1.38 lbs/MWh) where coal only accounts for 21.3% of the marginal generation. The R2 values are relatively high in MISO (0.74) and Central (0.73), which is consistent with the strong relationship between changes in coal-heavy generation and changes in SO2 emissions in MISO and Central. The R2 values are relatively low in the North (0.41) and South (0.42) sub-regions, this is because wind’s and gas’s significant contribution to marginal generation weakens the causal relationship between changes in generation and changes in SO2 emissions in North and South, respectively. The expanded MEFs for NOX is higher in MISO (1.30 lbs/MWh), North (1.43 lbs/MWh) and South (1.51 lbs/MWh) than those in Central (1.09 lbs/MWh). Similar to other emissions, the R2 value in the North (0.46), due to wind’s frequent presence in marginal generation, is the lowest compared to that in MISO (0.79), Central (0.71) and South (0.59). Table 1. The Expanded and Conventional Marginal Emission Factors (MEFs) for Regional (MISO) and Subregional (North, Central and South) Electricity Generation in 2014. There are significant differences in the average regional and sub-regional fuel mix: in MISO, the percentage of generation from coal, natural gas, nuclear, and wind is 57.7%, 15.5%, 16.2%, and 6.3%, respectively; in North, the percentage of generation from coal, natural gas, nuclear, and wind is 57.7%, 2.6%, 12.8%, and 22.5%, respectively; in Central, the percentage of generation from coal, natural gas, nuclear, and wind is 75.0%, 7.5%, 12.9%, and 2.4%, respectively; in South, the percentage of generation from coal, natural gas, and nuclear is 23.8%, 41.8%, and 26.1%, respectively. Region

Expanded MEFs

Conventional MEFs Expanded MEFs ~ Conventional MEFs diff.%

263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278

MISO North Central South MISO North Central South MISO North Central South

CO2 (tons/MWh) MEF ± 2 R2 0.76 ± 0.004 0.95 0.67 ± 0.010 0.64 0.82 ± 0.004 0.96 0.62 ± 0.004 0.93 0.78 ± 0.002 0.98 0.86 ± 0.004 0.95 0.85 ± 0.002 0.99 0.62 ± 0.004 0.93 2.8 28.1 3.7 0.0

SO2 (lbs/MWh) MEF ± 2 R2 2.59 ± 0.032 0.74 1.99 ± 0.052 0.41 3.34 ± 0.044 0.73 1.38 ± 0.034 0.42 2.67 ± 0.032 0.77 2.54 ± 0.044 0.61 3.47 ± 0.044 0.75 1.38 ± 0.034 0.42 3.1 28.4 3.9 0.0

NOX (lbs/MWh) MEF ± 2 R2 1.30 ± 0.014 0.79 1.43 ± 0.034 0.46 1.09 ± 0.014 0.71 1.51 ± 0.026 0.59 1.33 ± 0.014 0.82 1.87 ± 0.026 0.72 1.12 ± 0.014 0.73 1.51 ± 0.026 0.59 2.6 31.8 3.5 0.0

Marginal Fuel (%) coal gas wind 59.2 36.8 0.3 66.8 3.8 28.9 73.2 22.4 0.1 20.8 78.9 NA 61.4 37.2 NA 94.2 5.4 NA 76.6 22.8 NA 20.8 79.0 NA

Expanded MEFs ~ Conventional MEFs diff. % = (Conventional MEFs – Expanded MEFs)/Expanded MEFs × 100

3.2.Spatiotemporal Variations and Trends of Expanded MEFs We observe significant trends of the expanded MEFs across hours depending on the month. Figure 2 shows the seasonal and diurnal time variations of the expanded MEFs for CO2. In MISO, the expanded MEFs are higher, overall, at night than during the day. We attribute this to carbon-intensive coal generation that is often the primary contributor to meet the marginal demand in low-demand hours (see Figure 3), such as 11pm-2am next day. Moreover, coal-fired generators are less efficient during low-demand hours than average, because they have to operate at lower capacity factors.10 The expanded MEFs in the North sub-region are found to be lower in colder months, which is thought to be the result of greater availability of wind resources and consequently more wind generation on the margin. August has the highest expanded MEFs than any other month in the North, when wind resources are scarcest and the marginal demand for air conditioning is primarily fueled by coal. The expanded MEFs in Central do not have extreme contrasts across months and hours, but a few trends are still noticeable. For example, the 7 ACS Paragon Plus Environment

Environmental Science & Technology

Page 8 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297

expanded MEFs for CO2 during 10am-8pm in August and in the hour ending at 2pm in months from June to October are found to have lower MEFs for CO2 than neighboring hours. This difference is likely caused by marginal generation’s dependence on natural gas to meet peak demand for air conditioning during these time periods in this sub-region (see Figure 3). The expanded MEFs in South are relatively higher in the early morning than in the rest of the day particularly in the summer months, because marginal generation during low-demand hours is satisfied by coal- and gas-fired units (see Figure 3) and the coal-fired units’ ramp-up emission rates are usually high. The moderate expanded MEFs for CO2 in most hours and months reflect a largely constant level of gas-fired marginal generation in South, due to strong presence of natural gas in the system fuel mix. We also notice obvious differences of the expanded MEFs across MISO and its sub-regions (see SI section 3 and Figure S2). Overall, they vary the most in North and relatively less in Central and South. The outstanding low MEFs in North echo the fact that North has the greatest wind penetration in marginal generation compared to the other two sub-regions (see Table 1), while the relatively high MEFs in early-morning and late-night hours indicate the yet heavy reliance on coal in North. Central and South depend intensively on coal and natural gas for marginal generation, respectively, so the MEFs are higher in Central and relatively moderate in South. MISO

North

Central

South Expanded MEFs CO2 (tons/MWh)

1 Hour 0.5

0

298 299 300 301 302 303 304 305 306 307 308 309

Month

Month

Month

Month

Figure 2. Monthly and Hourly Expanded MEFs for CO2 for Regional (MISO) and Sub-regional (North, Central and South) Electricity Generation in 2014

3.3.Marginal Fuel Type, and MEFs Inspired by Figure 3 in Siler-Evans et al. (2012)7, we created Figure 3 to show the share of marginal generation by fuel type (top) and the expanded MEFs for CO2, SO2 and NOX (bottom) in correspondence to the level of system generation, which is a proxy for system demand. North and Central are both quite coal-heavy in generation. When system demand is low, coal is the dominant marginal fuel in both sub-regions, resulting in relatively high MEFs for CO2 (left axis) and SO2 (right axis). However, as demand increases, share of coal slightly decreases in North, but dramatically drops in Central. The share gap is gradually picked up by wind in North and 8 ACS Paragon Plus Environment

Page 9 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 310 311 312 313 314 315 316 317

318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340

quickly picked up by gas and hydro in Central. Such substitution of coal by wind and gas leads to the reduced MEFs for CO2 (left axis) and SO2 (right axis) in North and Central, respectively, but it has little effect on the MEFs for NOX level in both sub-regions. South is the most gasheavy sub-region. At the lowest demand level, coal accounts for roughly 55% of marginal generation, but declines to less than 2% at peak demand, while gas takes up the other 45% at low demand and almost all marginal generation at peak demand. The replacement of coal by gas in marginal demand causes the decline of MEFs for CO2 (left axis) and SO2 (right axis). The expanded MEFs for NOX rapidly increase with system demand and gas generation.

Figure 3. Share of Marginal Fuel Type (top) and Expanded MEFs for CO2, SO2 and NOX (bottom) as a Function of System Generation. Following from Siler-Evans et al. (2012)7: X axes are binned by fifth percentile of system generation (GWh) in each region and sub-region in 2014. Only coal (grey), gas (light blue), hydro (dark blue), other (yellow), and wind (green) are shown here because nuclear is rarely on the margin of generation. The expanded MEFs (bottom) have two Y axes: the left axis applies to CO2 (tons/MWh) and right axis applies to NOX and SO2 (lbs/MWh).

3.4.Application of MEFs We demonstrate several applications of the expanded MEFs for CO2 in this section. We examine bulk energy storage technologies that charge during low-demand hours (bottom 15%) and discharge during high-demand hours (top 15%) as well as demand side management (DSM) programs that are based on load shifting techniques and are designed to move demand load from high-demand hours to low-demand hours. These technologies are capable of shaving peak demand, maintaining grid reliability, and improving system efficiency.62,63 The bulk energy storage technologies are particularly useful in hedging against the variability of intermittent generation from renewable resources like wind and solar.42,43 Existing studies reveal that the operation of energy storage can cause emission increases.5,45,64–66 Hittinger and Azevedo (2015), for example, investigated and compared such increased effects across U.S. NERC regions. However, they used the conventional MEFs and neglected possible contribution of renewables like wind to marginal generation, which may overestimate the emission increases resulting from the operation of energy storage.5 In Table 2, we compare the expanded and conventional MEFs for CO2 during high-demand 9 ACS Paragon Plus Environment

Environmental Science & Technology

Page 10 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365

and low-demand hours for MISO and its sub-regions. The comparison illuminates that the conventional MEFs in general underestimate the potential emission increase (column “H ~ L diff. at 100% efficiency”) as compared with the expanded MEFs, even if the bulk energy storage technologies operate at a round-trip efficiency of 100%. This is especially true for the North subregion (0.168 versus 0.037). Although the expanded MEFs are lower compared to the conventional MEFs during both high-demand and low-demand hours in North, the increase of expanded MEFs from high-demand hours to low-demand hours is enlarged. This is because wind contributes to marginal generation more frequently during high-demand hours than during lowdemand hours (Figure 3). This consequently helps reduce the expanded MEFs CO2 more during high-demand hours than during low-demand hours. We also look at bulk energy storage technologies operating at a round-trip efficiency of 75%, which was indicated as the base-case in Hittinger and Azevedo (2015)5. The results show that the potential emission increase caused by operation of the storage technologies at 75% efficiency (column “H ~ L diff. at 75% efficiency” in Table 2) is significantly enlarged compared to that caused by the 100% efficient technologies. The comparison also reveals significant regional differences in the avoided CO2 emissions resulting from the same load-shifting intervention. The level of renewable penetration and the dependence on coal are the key factors that cause the differences across the regions. Table 2. Potential Emissions Increase Resulting from Bulk Energy Storage Technologies Charging during Lowdemand Hours and Discharging during High-demand Hours. The difference between MEFs in low-demand and high-demand hours is calculated using (Low-demand hours – High-demand hours). Values in parentheses are the percentage difference between MEFs in low-demand and high-demand hours, which is calculated using (Lowdemand hours – High-demand hours)/High-demand hours × 100. Region Expanded MEFs CO2 (tons/MWh) Conventional MEFs CO2 (tons/MWh)

366 367 368 369 370 371 372 373 374 375 376 377 378 379

MISO North Central South MISO North Central South

High-demand hours 0.622 0.601 0.649 0.559 0.673 0.857 0.735 0.559

Low-demand hours 0.811 0.769 0.881 0.700 0.826 0.894 0.884 0.700

H ~ L diff. at 100% efficiency 0.189 (30.3%) 0.168 (28.0%) 0.232 (35.7%) 0.141 (25.2%) 0.153 (22.7%) 0.037 (4.4%) 0.148 (20.2%) 0.141 (25.2%)

H ~ L diff. at 75% efficiency 0.459 (73.8%) 0.435 (70.7%) 0.525 (81.0%) 0.375 (67.0%) 0.428 (63.6%) 0.336 (39.2%) 0.443 (60.2%) 0.374 (66.9%)

4. Discussion It has been broadly acknowledged that marginal emission factors (MEFs) are more appropriate, than average emission factors, in evaluating the avoided emissions of interventions in the electricity system.6–8,27,67 The conventional MEFs focused on emitting sources such as coal and natural gas. This approach may suffice to assess the avoided emissions from interventions that aim to displace fossil-fueled generators 5,7,18 when there is low penetration and dispatchability of non-emitting sources such as wind. With renewables playing an increasingly important role and gaining improved dispatchability, however, the conventional MEFs may significantly overestimate the magnitude of avoided emissions and underestimate the role of renewables in shaping MEFs. Our expanded MEFs considering both emitting and non-emitting sources on the margin provides more accurate estimates of the avoided emissions from interventions including bulk energy storage, PHEVs, and demand response (DR). Our results have important implications for regional electricity system policy making. 10 ACS Paragon Plus Environment

Page 11 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416 417 418 419 420 421 422 423 424 425

Although the MISO overall has a low level of renewable penetration and renewables make a small contribution to marginal generation, there is significant heterogeneity at the sub-regional level. Wind penetration in the North is high and so is its contribution to marginal generation. Due to transmission congestions and renewables curtailment, further deployment of wind power in the North may not be as effective at reducing emissions as in other sub-regions of MISO that also have abundant wind resources. Energy policies using MEFs would benefit from specified consideration of sub-regional differences in fuel mix. We find significant hour-to-hour differences of the expanded MEFs for CO2 in North, but modest hour-to-hour differences in Central and South. In North, the MEFs for CO2 tend to be lower in winter and nighttime when system demand is low and the wind resource is abundant and more often on the margin; whereas in Central and South, the MEFs for CO2 tend to be higher during night hours, when system demand is low and coal and gas are more often on the margin. The spatiotemporal heterogeneity of MEFs provide valuable guidance for policy makers and practitioners to evaluate the environmental impacts of emerging technologies with respect to their operational characteristics. For example, PHEVs that are scheduled to charge overnight will end up having quite low emission impacts if their owners reside in North, but relatively high emission impacts if their owners live in Central and South. We note that deploying bulk energy storage technologies and demand-side management programs in MISO to shift 1 MWh of electricity generation from high- to low-demand hours can result in a CO2 emission increase of 0.189 tons or 30%, if the technologies have a round-trip efficiency of 100%, and 0.459 tons or 73.8%, if the technologies have a more realistic round-trip efficiency of 75%. The emission shortcoming of storage technologies has been recognized in previous studies and may jeopardize many valuable services provide by the storage technologies, including reliability, responsiveness, and integration of intermittent renewable resources.5,45,64 However, the studies underestimate the potential emission increases due to neglecting the marginal generation from renewables.5,45,64 We find that, in the North sub-region where wind accounts for almost one third of the marginal generation, the negligence leads to underestimating the marginal emission increase from storage technologies by 4.5 times, if the technologies have a round-trip efficiency of 100%, and 1.3 times, if the technologies have a more realistic round-trip efficiency of 75%. Therefore, policy makers and technology investors ought to be cognizant of the issues when assessing the value of additional bulk energy storage. Our expanded MEFs have accounted for MISO system-wide generation profile changes in recent years and comprehensively estimate the avoided emissions of interventions in the current MISO electricity system. We recommend other markets of Independent System Operators (ISO) and Regional Transmissions Operators (RTO) to consider the impact of renewables when calculating MEFs. Our method and estimates may be further improved if MISO discloses the locational marginal generators (LMG) that are dispatched to meet the next increment of system load and balance the system. Future improvement in MEFs methods will enable proper assessment of policy and technology interventions in terms of their societal impacts, and will stimulate effective investment and policy decisions. A limitation of our study is the omission of electricity trade in estimation of MEFs. Detailed data on how electricity flows between MISO sub-regions or between MISO and other regions are not readily available. Electricity import rate in different eGRID sub-regions ranges from 0 to 30%, although 30% would be an extreme case and 15% is already an upper bound for many regions;68 monthly electricity net interchange rate between MISO and other balancing authority regions ranges from 2 to 12%.69 Assuming an unlikely net interchange rate of 30% between 11 ACS Paragon Plus Environment

Environmental Science & Technology

Page 12 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 426 427 428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465 466 467 468 469 470 471

MISO sub-regions, the expanded MEFs would be lower than the conventional MEFs by about 10% in the wind-rich North sub-region (Table S2 and Figure S3). And assuming a more reasonable 15%, the expanded MEFs would be lower by 14-18%. The results indicate that considering electricity interchange would not affect our main message that it is important to incorporate renewables in MEFs estimation in high renewable penetration regions so as to provide more accurate estimates of avoided emissions by policy interventions. This is especially true given the trend of continuous renewables expansion around the world. Associated content Supporting Information. Map of MISO and its sub-regions, total generation and fuel mix in 2014 by MISO and its subregions, linear regression details of the expanded MEFs, hourly trends of expanded and conventional MEFs in 2014 for MISO and its sub-regions (year-round, summer, and winter), and sensitivity analysis on the MEFs based on simulated sub-regional electricity net interchange rates. Acknowledgements The authors acknowledge the generous support of the University of Minnesota’s MN-DRIVE program and the Institute on the Environment. We also would like to thank Jennifer Schmitt and Mo Sun for comments and suggestions. Reference (1) Climate change 2014: synthesis report; Pachauri, R. K., Mayer, L., IPCC, Eds.; Intergovernmental Panel on Climate Change: Geneva, Switzerland, 2015. (2) Wang, S.; Hao, J. Air quality management in China: Issues, challenges, and options. J. Environ. Sci. 2012, 24 (1), 2–13. (3) Smith, K. R.; Jerrett, M.; Anderson, H. R.; Burnett, R. T.; Stone, V.; Derwent, R.; Atkinson, R. W.; Cohen, A.; Shonkoff, S. B.; Krewski, D.; et al. Public health benefits of strategies to reduce greenhouse-gas emissions: health implications of short-lived greenhouse pollutants. The lancet 2010, 374 (9707), 2091–2103. (4) Lutsey, N.; Sperling, D. America’s bottom-up climate change mitigation policy. Energy Policy 2008, 36 (2), 673–685. (5) Hittinger, E. S.; Azevedo, I. M. L. Bulk Energy Storage Increases United States Electricity System Emissions. Environ. Sci. Technol. 2015, 49 (5), 3203–3210. (6) Hawkes, A. D. Estimating marginal CO2 emissions rates for national electricity systems. Energy Policy 2010, 38 (10), 5977–5987. (7) Siler-Evans, K.; Azevedo, I. L.; Morgan, M. G. Marginal Emissions Factors for the U.S. Electricity System. Environ. Sci. Technol. 2012, 46 (9), 4742–4748. (8) Doucette, R. T.; McCulloch, M. D. Modeling the CO2 emissions from battery electric vehicles given the power generation mixes of different countries. Energy Policy 2011, 39 (2), 803–811. (9) Hawkes, A. D. Long-run marginal CO2 emissions factors in national electricity systems. Appl. Energy 2014, 125, 197–205. (10) Ryan, N. A.; Johnson, J. X.; Keoleian, G. A. Comparative Assessment of Models and Methods To Calculate Grid Electricity Emissions. Environ. Sci. Technol. 2016, 50 (17), 8937–8953. 12 ACS Paragon Plus Environment

Page 13 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516

(11) Peterson, S. B.; Whitacre, J. F.; Apt, J. Net Air Emissions from Electric Vehicles: The Effect of Carbon Price and Charging Strategies. Environ. Sci. Technol. 2011, 45 (5), 1792– 1797. (12) McCarthy, R.; Yang, C. Determining marginal electricity for near-term plug-in and fuel cell vehicle demands in California: Impacts on vehicle greenhouse gas emissions. J. Power Sources 2010, 195 (7), 2099–2109. (13) Axsen, J.; Kurani, K. S.; McCarthy, R.; Yang, C. Plug-in hybrid vehicle GHG impacts in California: Integrating consumer-informed recharge profiles with an electricity-dispatch model. Energy Policy 2011, 39 (3), 1617–1629. (14) Blumsack, S.; Samaras, C.; Hines, P. Long-term electric system investments to support plug-in hybrid electric vehicles. In Power and Energy Society General Meeting-Conversion and Delivery of Electrical Energy in the 21st Century, 2008 IEEE; IEEE, 2008; pp 1–6. (15) Kintner-Meyer, M.; Schneider, K.; Pratt, R.; Pacific Northwest National Laboratory. Impacts Assessment of Plug-In Hybrid Vehicles on Electric Utilities and Regional U.S. Power Grids Part 1: Technical Analysis. (16) Holland, S. P.; Mansur, E. T. Is real-time pricing green? The environmental impacts of electricity demand variance. Rev. Econ. Stat. 2008, 90 (3), 550–561. (17) Archsmith, J.; Kendall, A.; Rapson, D. From cradle to junkyard: assessing the life cycle greenhouse gas benefits of electric vehicles. Res. Transp. Econ. 2015, 52, 72–90. (18) Graff Zivin, J. S.; Kotchen, M. J.; Mansur, E. T. Spatial and temporal heterogeneity of marginal emissions: Implications for electric cars and other electricity-shifting policies. J. Econ. Behav. Organ. 2014, 107, 248–268. (19) Jansen, K. H.; Brown, T. M.; Samuelsen, G. S. Emissions impacts of plug-in hybrid electric vehicle deployment on the U.S. western grid. J. Power Sources 2010, 195 (16), 5409–5416. (20) Stephan, C. H.; Sullivan, J. Environmental and Energy Implications of Plug-In HybridElectric Vehicles. Environ. Sci. Technol. 2008, 42 (4), 1185–1190. (21) Tamayao, M.-A. M.; Michalek, J. J.; Hendrickson, C.; Azevedo, I. M. L. Regional Variability and Uncertainty of Electric Vehicle Life Cycle CO 2 Emissions across the United States. Environ. Sci. Technol. 2015, 49 (14), 8844–8855. (22) Yuksel, T.; Tamayao, M.-A. M.; Hendrickson, C.; Azevedo, I. M. L.; Michalek, J. J. Effect of regional grid mix, driving patterns and climate on the comparative carbon footprint of gasoline and plug-in electric vehicles in the United States. Environ. Res. Lett. 2016, 11 (4), 044007. (23) Holland, S. P.; Mansur, E. T.; Muller, N. Z.; Yates, A. J. Environmental benefits from driving electric vehicles?; National Bureau of Economic Research, 2015. (24) Holland, S. P.; Mansur, E. T.; Muller, N. Z.; Yates, A. J. Distributional Effects of Air Pollution from Electric Vehicle Adoption; National Bureau of Economic Research, 2016. (25) Diem, A.; Quiroz, C. How to use eGRID for Carbon Footprinting Electricity Purchases in Greenhouse Gas Emission Inventories; U.S. EPA, 2012. (26) U.S. EPA. Clean Power Plan: Description of 2012 Unit - level Data Using the eGRID Methodology https://www.epa.gov/sites/production/files/2014-06/documents/20140602description-egrid-methodology.pdf (accessed Nov 1, 2016). (27) Marnay, C.; Fisher, D.; Murtishaw, S.; Phadke, A.; Price, L.; Sathaye, J. Estimating carbon dioxide emissions factors for the California electric power sector. Lawrence Berkeley Natl. Lab. 2002.

13 ACS Paragon Plus Environment

Environmental Science & Technology

Page 14 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 517 518 519 520 521 522 523 524 525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561

(28) Raichur, V.; Callaway, D. S.; Skerlos, S. J. Estimating Emissions from Electricity Generation Using Electricity Dispatch Models: The Importance of System Operating Constraints. J. Ind. Ecol. 2016, 20 (1), 42–53. (29) Newcomer, A.; Blumsack, S. A.; Apt, J.; Lave, L. B.; Morgan, M. G. Short Run Effects of a Price on Carbon Dioxide Emissions from U.S. Electric Generators. Environ. Sci. Technol. 2008, 42 (9), 3139–3144. (30) Bettle, R.; Pout, C. H.; Hitchin, E. R. Interactions between electricity-saving measures and carbon emissions from power generation in England and Wales. Energy Policy 2006, 34 (18), 3434–3446. (31) Voorspools, K. R.; D D’haeseleer, W. An evaluation method for calculating the emission responsibility of specific electric applications. Energy Policy 2000, 28 (13), 967–980. (32) Kim, J. D.; Rahimi, M. Future energy loads for a large-scale adoption of electric vehicles in the city of Los Angeles: Impacts on greenhouse gas (GHG) emissions. Energy Policy 2014, 73, 620–630. (33) Choi, D. G.; Kreikebaum, F.; Thomas, V. M.; Divan, D. Coordinated EV Adoption: Double-Digit Reductions in Emissions and Fuel Use for $40/Vehicle-Year. Environ. Sci. Technol. 2013, 130829161516000. (34) van Vliet, O.; Brouwer, A. S.; Kuramochi, T.; van den Broek, M.; Faaij, A. Energy use, cost and CO2 emissions of electric cars. J. Power Sources 2011, 196 (4), 2298–2310. (35) Lund, H.; Kempton, W. Integration of renewable energy into the transport and electricity sectors through V2G. Energy Policy 2008, 36 (9), 3578–3587. (36) Yuksel, T.; Michalek, J. J. Effects of Regional Temperature on Electric Vehicle Efficiency, Range, and Emissions in the United States. Environ. Sci. Technol. 2015, 49 (6), 3974–3980. (37) McConnell, B. W.; Hadley, S. W.; Xu, Y. A Review of Barriers to and Opportunities for the Integration of Renewable Energy in the Southeast; ORNL/TM-2010/138; Oak Ridge National Laboratory: Oak Ridge, TN, 2011. (38) Keith, G.; Biewald, B.; Sommer, A.; Henn, P.; Breceda, M. Estimating the emission reduction benefits of renewable electricity and energy efficiency in North America: experience and methods. Energy 2003, 5255, 5264–7181. (39) Hadley, S. W.; Van Dyke, J. W. Emissions benefits of distributed generation in the Texas market; ORNL/TM-2003/100; Oak Ridge National Laboratory: Oak Ridge, TN, 2003. (40) Wang, K.; Wei, Y.-M. China’s regional industrial energy efficiency and carbon emissions abatement costs. Appl. Energy 2014, 130, 617–631. (41) Du, L.; Mao, J. Estimating the environmental efficiency and marginal CO2 abatement cost of coal-fired power plants in China. Energy Policy 2015, 85, 347–356. (42) Anderson, D.; Leach, M. Harvesting and redistributing renewable energy: on the role of gas and electricity grids to overcome intermittency through the generation and storage of hydrogen. Energy Policy 2004, 32 (14), 1603–1614. (43) Kanoria, Y.; Montanari, A.; Tse, D.; Zhang, B. Distributed storage for intermittent energy sources: Control design and performance limits. In Communication, Control, and Computing (Allerton), 2011 49th Annual Allerton Conference on; IEEE, 2011; pp 1310– 1317. (44) Stadler, M.; Siddiqui, A.; Marnay, C.; Aki, H.; Lai, J. Control of greenhouse gas emissions by optimal DER technology investment and energy management in zero-net-energy buildings. Eur. Trans. Electr. Power 2011, 21 (2), 1291–1309.

14 ACS Paragon Plus Environment

Page 15 of 17

Environmental Science & Technology

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 562 563 564 565 566 567 568 569 570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605

(45) Carson, R. T.; Novan, K. The private and social economics of bulk electricity storage. J. Environ. Econ. Manag. 2013, 66 (3), 404–423. (46) Newcomer, A.; Apt, J. Near-Term Implications of a Ban on New Coal-Fired Power Plants in the United States. Environ. Sci. Technol. 2009, 43 (11), 3995–4001. (47) MISO. Monthly Market Report 2014 https://www.misoenergy.org/Library/Pages/ManagedFileSet.aspx?SetId=2070 (accessed Nov 1, 2016). (48) PJM. Marginal Fuel Type Data http://www.pjm.com/markets-and-operations/energy/realtime/historical-bid-data/marg-fuel-type-data.aspx (accessed Nov 1, 2016). (49) MISO. MISO 2011 Wind Generation Update - With the focus on Dispatchable Intermittent Resources, 2011. (50) Stafford, B. A.; Wilson, E. J. Winds of change in energy systems: Policy implementation, technology deployment, and regional transmission organizations. Energy Res. Soc. Sci. 2016, 21, 222–236. (51) MISO. Wind Forecasting Review, 2015. (52) MISO. MISO Informational Forum, 2015. (53) ISO New England. 2014 ISO New England Electric Generator Air Emissions Report; 2016. (54) MISO. MISO 2014 Real-Time Fuel on the Margin Report https://www.misoenergy.org/Library/Repository/Market%20Reports/2014_rt_fuel_on_mar gin.zip (accessed Nov 5, 2016). (55) MISO. MISO 2015 Real-Time Fuel on the Margin Report https://www.misoenergy.org/Library/Repository/Market%20Reports/2015_rt_fuel_on_mar gin.zip (accessed Nov 5, 2016). (56) MISO. MISO 2016 Real-Time Fuel on the Margin Report https://www.misoenergy.org/Library/Repository/Market%20Reports/2016_rt_fuel_on_mar gin.zip (accessed Nov 5, 2016). (57) MISO. MTEP15 Chapter 9.1: MISO Overview; 2015. (58) MISO. MISO Historical Generation Fuel Mix for 2014 https://www.misoenergy.org/RTDisplays/MKTRPT/AllReportsList.html?rptName=Historic al%20Generation%20Fuel%20Mix%20for%202014%20(xls) (accessed Nov 5, 2016). (59) U.S. EPA. Air Markets Program Data | Clean Air Markets | US EPA https://ampd.epa.gov/ampd/ (accessed Nov 5, 2016). (60) EIA. Annual Electric Generator data - EIA-860 data file http://www.eia.gov/electricity/data/eia860/ (accessed Nov 10, 2016). (61) MISO. MISO Fuel Mix Report Readers Guide.pdf. 2014. (62) Moura, P. S.; de Almeida, A. T. The role of demand-side management in the grid integration of wind power. Appl. Energy 2010, 87 (8), 2581–2588. (63) Wang, P.; Huang, J. Y.; Ding, Y.; Loh, P.; Goel, L. Demand side load management of smart grids using intelligent trading/metering/ billing system. In 2011 IEEE Trondheim PowerTech; 2011; pp 1–6. (64) Lueken, R.; Apt, J. The effects of bulk electricity storage on the PJM market. Energy Syst. 2014, 5 (4), 677–704. (65) Hittinger, E.; Whitacre, J. F.; Apt, J. Compensating for wind variability using co-located natural gas generation and energy storage. Energy Syst. 2010, 1 (4), 417–439.

15 ACS Paragon Plus Environment

Environmental Science & Technology

Page 16 of 17

THIS IS A DRAFT. DO NOT DISTRIBUTE WITHOUT PERMISSION FROM AUTHORS 606 607 608 609 610 611 612 613 614 615 616 617

(66) Denholm, P.; Kulcinski, G. L. Life cycle energy requirements and greenhouse gas emissions from large scale energy storage systems. Energy Convers. Manag. 2004, 45 (13– 14), 2153–2172. (67) Farhat, A. A. M.; Ugursal, V. I. Greenhouse gas emission intensity factors for marginal electricity generation in Canada. Int. J. Energy Res. 2010, 34 (15), 1309–1327. (68) Diem, A.; Quiroz, C. How to Use eGRID for Carbon Footprinting Electricity Purchases in Greenhouse Gas Emission Inventories. Environ. Prot. Agency Httpswww Epa Govsitesproductionfiles2015-01documentsadiem Pdf 2012. (69) EIA. U.S. Electric System Operating Data https://www.eia.gov/beta/realtime_grid/#/data/table?end=20161030&start=20160930&freq uency=Daily&dataTypes=mc&bas=00000000000g (accessed Jul 14, 2017).

16 ACS Paragon Plus Environment

Percentage of Time a Fuel is on the Margin in MISO

100%

PageEnvironmental 17 of 17 Science & Technology 80% 60% 40% 20%

ACS Paragon Plus Environment

0%

Coal Natural Hydro Nuclear Other Wind Gas